Vehicle and Person Monitoring - Commute Warrior Android App Randall Guensler School of Civil and Environ. Engineering Georgia Institute of Technology Atlanta Regional Commission Model Users Group (MUG) February 28, 2014
Background Since 1996, Georgia Tech researchers have been developing a wide variety of technologies for monitoring travel behavior and vehicle activity Vehicle instrumentation Android App cellphone systems Researchers have also developed server-side technologies for data handling and analytical work Fleet management Vehicle tracking and event reporting Server-side data processing All-electronic travel diaries
Commute Atlanta Diary vs. GPS Comparison 29% of trips not reported in diaries (670 of 2,292 trips) 5.9 trips/day/household (reported by participants) 8.4 trips/day/household (recorded by GPS) Under-reporting of VMT by 22% More short distance/duration trips go unreported Only 24% of households reported accurately 13% of households yield 50% of unreported trips Under-reporting is not random Demographic impacts Temporal impacts
Diary Comparison Conclusions Trip item non-response is significant Travel during daytime work hours, and commercial use of vehicles significantly impacts reporting probability Trips made after the Return to Home late in the day make up a significant portion of unreported trips Change retrieval process ask for last trip of each day and work toward that trip to help reduce non-response Use of instrumented vehicles to supplement or replace diaries holds promise for much more accurate data
Fleet Tracking System Minneapolis Deployment GPS tracking Cellular reporting Second-by-second real-time data On/Off ports Monitor alarms Trigger events Serial port I/O
Real-time Vehicle Tracking Last 2 minutes of Operations
Trip Purpose Data Collection Real-time vehicle tracking Internet travel diaries Collect trip purpose information on all activities undertaken at each stop In-vehicle systems
Online Electronic Travel Diaries Minneapolis demonstration 46 participants, 6-14 travel days requested 85% diary completion rate, all fields 200% data provision rate (extra days reported)
Commute Warrior Android App
Commute Warrior Android App for personal cell phone tracking Monitors second-by-second phone location and speed Transmits real-time data to server via cellular Settings for Wi-Fi only transmission if desired Integrates the functionality of previous vehicle monitoring systems, including travel diary reporting Runs in background with low battery power draw Final beta tests November/December 2013 Public deployment of 200 units March 2014 GRTA express bus passenger volunteers
Commute Warrior Android App App runs in background Always on Can stop recording at any time (via menu) for privacy Commute Warrior Widget Shows status (running) Shows data collection and transmission activity Links to main App window
Commute Warrior Main Interface Settings Menu (gear icon) Travel Journal Stop Recording Hide Window Runs in background Automatically restarts
Commute Warrior Travel Journal Calendar Interface All days for which trip data were recorded appear in green Page back/forward by month Select any day to view trip journal
Commute Warrior Trip Viewer and Diary Interface Daily record of all trips Includes: Trip date/time Trip duration Trip distance Map icon (left) for trip playback Survey icon (right) to record trip purpose data Trip purpose survey (short) Data links to long-form Internet-based survey
Commute Warrior Trip Viewer Facilitates trip purpose recall Open Street Map (OSM) Pan and zoom features Center and auto-center Slider bar for trip playback Flags - green (start), red (end) Trip playback animation Red icon moves with travel Visualize speed and delays
Commute Warrior Travel Diary Customizable Trip Data Entry Primary trip purpose Drop down box Applicable business activities Short form for App entry Long-form via Web interface Includes more complex nested options such as specific customer locations
Commute Warrior Travel Diary Survey Requests Users complete diary entries for any trips desired any time Server is programmed to issue diary-day requests Regular (e.g., Mondays) Random (e.g., 7x/month) Targeted days Special events Interesting data event identified by analysts
Trip Purpose Approach Primary trip purpose The primary activity motivating the travel Secondary trip purposes Additional activities undertaken during travel Trip place data entry for each primary/secondary purpose
Automation of Data Analysis Vehicle locations processed in real-time Real-time data visualization (mapping functions) Data structures are amenable to automated creation of summary charts, tables, and text reports Summarize data at any aggregation level, from individual trip to total annual activity Data summaries are customer-accessible Web portal for access to data, reports, and maps Reports can be automatically e-mailed
Maps, Reports, and Data are Available through the Web Portal Activity Monitoring Current position maps
Website Interface Fleet managers and drivers can access information associated with every trip via a website Login features allow managers to control access to travel data for individual vehicles and groups of vehicles Calendar functions allow users to examine trips by date Monthly vehicle travel summaries, activity logs, daily driver reports, and a wide-variety of automated reports are tailored for each deployment and uploaded to the site Drivers and fleet managers can easily update driver and vehicle data using online forms
Website Capabilities Automatic daily, weekly, or monthly reports Email, text, or call each bus driver Individual summaries of activity can be sent to drivers on a per-bus basis via email, so that drivers can keep track of their personal activity Drivers can receive information about their personal performance, and comparisons to other anonymous drivers across schools or fleet-wide Past information about any vehicle trip on any day Accessible to company via web interface system
Additional Reporting Safety reports Speed vs. speed limit Seatbelt usage Hard braking and acceleration events Emissions and idling Route adherence Schedule adherence Special requests Web portal access to automated reports (by vehicle, route, time period, etc.)
Speed Histograms Three Commute Atlanta Participants 20% 18% Age 19 Single Female 2-Door Coupe Percentage 16% 14% 12% 10% 8% 6% 4% 2% % Operation 0% 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90+ Speed 20% 18% Age 30 Single Female 4-Door Sedan Percentage 16% 14% 12% 10% 8% 6% 4% % Operation 2% 0% 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90+ Speed Age 59 Married Female 2 Children 4-Door Sedan Percentage 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% % Operation 0% 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90+ Speed 70 mph Speed in 5mph Bins (0 to 90+)
Activity Space for Four Vehicles Two Months of Travel Vehicle 1: HH Size = 3, Age = 28 Vehicle 2: HH Size = 3, Age = 29 Vehicle 3: HH Size = 3, Age = 52 Vehicle 4: HH Size = 2, Age = 70, Retired
Interactive Speed Maps
Geo-fencing Applications
Mapping Bus Fleet Idling Events Idle events in Cobb County Categorized using GIS geo-fencing and proximity to certain features Size of circle represents duration of the event
Human Subjects Issues Electronic monitoring captures personally-identifiable driver behavior, vehicle activity, and travel patterns Some data can be detrimental if disclosed (e.g. trip location and time, speeding, seat belt use, etc.) Federal law, federal regulations, state law, and state regulations apply to the data collection Potential legal liability for disclosure of data Privacy and security of data are paramount and must be formally addressed
Conclusions Technology foundation developed at Georgia Tech New server and data/mapping systems are ready to support major fleet deployments Extensive experience in automating complex data processing and analysis Commute Warrior App Deployment in March 2014 with 200 users Deployment in Summer 2014 with 2000 users National Center for Sustainable Transportation UTC research should lead to large scale 2015 deployments